{"id":"5609c5a9-f530-436a-bd69-e4f2acd7f526","arxiv_id":"2501.15756","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Introduces a piecewise-linear flow on cluster complexes whose leaves generalize green mutation, proving these complexes are spheres for Dynkin quivers and contractible for Euclidean quivers.","lead":"The paper constructs a continuous flow on cluster complexes, called X-evolution, which sweeps from one singular point to another. It shows this flow generalizes green mutation and recovers the topology of cluster complexes for Dynkin and Euclidean quivers.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 3.12's proof of well-definedness fails when c<0: the point Y in (3.18) can have negative barycentric coordinates, so the X-evolution flow is not rigorously defined for arbitrary points X.","rationale":"The reader's weakest assumption was Condition (*), the finiteness condition guaranteeing that X-evolving triangles exist and are unique. My concern is different: even under Condition (*), the proof of Lemma 3.12 contains a false assertion in the c<0 case, so the well-definedness of the evolution flow is not established for arbitrary points X. Since Theorem 4.2 and all its consequences depend on this lemma, the paper needs a nontrivial proof fix before the central claim can be accepted as proven. However, the flaw is localized and likely repairable by a tangent-cone argument: at an interior point V of a cell, each summand in (3.13) lies in the tangent cone of a cell containing V, so the sum may well be tangent even when Y is not a point. The paper's explicit examples and the overall construction are consistent with the theorem being true; the issue is the rigor of the core well-definedness step, not an evident counterexample. Thus the verdict should remain conditional rather than outright rejection, but the condition is stronger than the reader stated: the authors must supply a correct proof of Lemma 3.12 for all signs of c.","tokens_in":40394,"tokens_out":18422,"duration_ms":179103,"concrete_test":"In the A2 cluster category with the labelling above, compute explicitly the downward X-evolving triangles for X = 0.2 a_0 + 0.8 a_2 and the cell V = {a_0, a_3}, with V = 0.5 a_0 + 0.5 a_3. Verify: (1) that Y from (3.18) is -1/3 a_0 + 4/3 a_3, hence not a point; (2) that Evo_Ó_X(V) from (3.13) nevertheless lies in the tangent cone of Cpx(A2) at V, so V + epsilon Evo remains in the complex for small epsilon. If (2) holds, the theorem may survive with a repaired proof of Lemma 3.12; if (2) fails, Theorem 4.2 is false.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Theorem 4.2 rests on Lemma 3.12, which claims that the X-evolution flow is well-defined for every point X. In the proof, after the partition I0, I1, J and the scalar c = sum_{i0 in I0} c_{i0} - sum_{i1 in I1} c_{i1}, the case c != 0 is handled by writing Evo_Ó_X(V) = c(Y - V), where Y is claimed to be a point of the cell Y from (3.15). This claim is false when c < 0. Then Y = (P - N)/c with P = sum_{i0} c_{i0} X_{i0} + sum_J c_j W_j and N = sum_J c_j U_j + sum_{i1} c_{i1} X_{i1}[1]. Since P and N are convex combinations, their difference divided by a negative scalar generally has negative coordinates. Explicitly, in the A2 cluster category, write the five indecomposables as a_0,...,a_4 with tau a_i = a_{i+1} and compatible pairs {a_i, a_{i+2}}, {a_i, a_{i+3}}. Take X = 0.2 a_0 + 0.8 a_2 and V = 0.5 a_0 + 0.5 a_3 (interior of the edge {a_0,a_3}). Then I0 = {0}, I1 = {2}, c = -0.6, and (3.18) gives Y = -1/3 a_0 + 4/3 a_3, not a point. Thus the written proof of Lemma 3.12 does not establish that V + t Evo_Ó_X(V) stays in the cluster complex for small t>0. A separate tangent-cone argument may repair this, but it is absent; as written, the foundational local flow is not proven for all points X.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces, for any point X in the cluster complex Cpx(C) of a 2-Calabi-Yau category C, an X-evolution flow and an associated one-dimensional piecewise linear X-foliation with singularities X and X[1]. It further claims that these flows refine Keller's green mutation, that for Dynkin and Euclidean quivers the foliations are compact or semi-compact for suitable X, and that these facts imply known homotopy-type results for cluster complexes and the generation of the fundamental group of the cluster exchange graph by squares and pentagons. The main technical steps are the construction of downward/upward X-evolving triangles, a family of flows for arbitrary points obtained by linear combination, the gluing of local leaves into a global foliation, and interval-decreasing arguments in Auslander–Reiten theory in the Dynkin/Euclidean cases.","tokens_in":40799,"tokens_out":5805,"duration_ms":54659,"significance":"If the main results are correct, the paper provides a genuinely new continuous dynamical structure on cluster complexes, unifying discrete mutation phenomena and giving a uniform route to the homotopy type of cluster complexes in the Dynkin and Euclidean cases. The categorical framework is carefully set up, the statements are precise, and the paper contains many detailed examples (A2, A3, D4, affine A) that illustrate the constructions. The application to fundamental groups generated by squares and pentagons is a natural and attractive consequence. However, the central construction of the flow for an arbitrary point X is not rigorously established as written, and since Theorem 4.2 depends directly on that construction, the paper cannot be accepted in its present form.","major_comments":[{"comment":"The proof of well-definedness of Evo^Ó_X for arbitrary points X is not valid. In (3.18), the claimed point Y is defined as Y = (P - N)/c, where P is a convex combination of the X_{i0} and W_j and N is a convex combination of the U_j and X_{i1}[1]. The coefficients of Y sum to 1, but they are not all nonnegative: when c > 0 the coefficients of the U_j and X_{i1}[1] terms are negative, and when c < 0 the coefficients of the X_{i0} and W_j terms are negative. Therefore Y is generally an affine combination, not a point of the simplex (3.15). For a concrete instance, in the A2 cluster category write the five indecomposables as a_0,...,a_4 with compatible pairs {a_i,a_{i+2}} and {a_i,a_{i+3}}, take X = 0.2a_0 + 0.8a_2 and V = 0.5a_0 + 0.5a_3; then I0 = {0}, I1 = {2}, c = -0.6, and (3.18) gives Y = -1/3 a_0 + 4/3 a_3, which is not a point of the cluster complex. Consequently the proof does not show that V + t Evo^Ó_X(V) remains in Cpx(C) for small t > 0. Since Lemma 3.12 is the foundation for the local flow used in Theorem 4.2 and Lemma 4.1, the central foliation theorem is not proven. The argument may be repairable by a separate tangent-cone or controlled-trajectory argument, but such an argument is absent.","section":"Section 3.3, Lemma 3.12, Eq. (3.18)"},{"comment":"The proof of Theorem 4.2 is only one short paragraph and relies entirely on Lemma 3.12 for the definition of the flow and on Lemma 4.1 for the gluing of local leaves. Because Lemma 3.12 is not established for all points, the conclusion that 'the X-evolution flow induces a piecewise linear foliation' is not justified. The statement of Theorem 4.2 may be true, but the present proof is incomplete at a load-bearing point. Please either repair Lemma 3.12 or reformulate Theorem 4.2 under hypotheses for which well-definedness is proven.","section":"Section 4.1, Theorem 4.2"},{"comment":"In the proof of compactness for Euclidean quivers with X a rigid regular simple, the assertion 'It is similar to 1° that 3° holds' is too terse. Lemma B.4(3°) is a key step in the interval-decreasing argument that shows leaves enter Starp(X), and the claimed similarity is not immediate because the roles of the two sections and the vector-bundle interval are different. Please expand this step so that the reader can verify it without reconstructing the entire tube geometry.","section":"Appendix B.2, Lemma B.4(3°)"}],"minor_comments":[{"comment":"The phrase 'clique complex for the compatibility relation on the ground set' should specify whether the ground set is the set of isomorphism classes of rigid indecomposables or a chosen representative set; this matters for the realization of points such as X = sum c_i X_i.","section":"Section 2.4, Definition 2.5"},{"comment":"In (3.5), the symbols W and U are used both for objects and for points of the cluster complex associated with those objects; please make the convention explicit, as in Convention 3.1, to avoid confusion in formulas such as W - V.","section":"Section 3.2, Definition 3.9"},{"comment":"The word 'proofed' appears twice in the proof of the second set of equivalences; it should be 'proved'.","section":"Section 3.1, Definition/Lemma 3.2"},{"comment":"There is a typo 'Grothedieck' for 'Grothendieck' in the sentence about K_0 groups.","section":"Section 5.2"},{"comment":"Several reference entries contain typographical errors, including 'combinactorics', 'Cambrigde', 'djoimension', and 'isomorphc'; a careful proofreading of the bibliography is recommended.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The main flaw is localized to Lemma 3.12, but it blocks the central theorem; I do not see it as an irreparable error. The paper draws heavily on the first author's prior work, and the exposition in Section 5 and Appendix B assumes familiarity with that work; the editor may wish to ask the authors to make the independence of the new results from the prior tools clearer. The examples and the overall program are promising, and with a corrected proof of Lemma 3.12 the paper could be suitable for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This paper introduces a genuinely new continuous structure for cluster complexes: for any point X in Cpx(C), a flow whose only singularities are X and X[1], and a piecewise-linear foliation. The main new content is the construction itself, Theorem 4.2, and Theorem 5.10, which identifies the orientation induced by the flow on the exchange graph with green mutation for generic X. The applications to Dynkin/Euclidean homotopy type and to generation of pi_1(CEG) by squares and pentagons are real, and the spherical/contractible results reproduce known facts in a uniform way.\n\nI checked the stress-test concern about Lemma 3.12 and the case c<0. I think it misses the mark. The quantity Y in (3.18) is not required to be a point of the simplex; it is an affine combination in the span of the bigger cell. Since the point V is in the interior of its own cell, the tangent cone is the full affine span, so for any nonzero direction d the segment V+td stays inside for small t. The proof would be clearer if it said 'affine coordinates' instead of 'point', and the reuse of the letter Y in Lemma 4.1 is confusing, but the vector identity (3.17) is all that is used. This is an expositional trap, not a load-bearing flaw.\n\nThe real soft spots are elsewhere. The condition (*) at the start of Section 3 is fine for cluster categories of Jacobi-finite non-degenerate QPs but not automatic, and the main theorems inherit it. The Dynkin/Euclidean proofs in Appendix B are technical and somewhat compressed; the interval-decreasing lemmas and the AR-theory identifications need a careful referee. The abstract also overstates Theorem 5.10 by dropping the genericity condition on X — the theorem itself is precise, so this is fixable. There is heavy self-citation for background, but the cited tools have independent published proofs; I do not see circularity.\n\nWho this is for: anyone working on cluster complexes, exchange graphs, or continuous models of cluster combinatorics. It deserves a serious referee, not a desk reject. I would send it with a request to expand the appendix proofs and to adjust the abstract.","headline":"A genuinely new continuous flow/foliation on cluster complexes, with a solid core and compressed Dynkin/Euclidean proofs; the stress-test concern about Lemma 3.12 does not land.","tokens_in":41306,"tokens_out":8631,"would_cite":true,"duration_ms":89929,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["13F60","16G20","16G70","18G80"],"pacs":[],"model":"deepseek-v4-flash","headline":"For every point X in a cluster complex, the X-evolution flow defines a piecewise linear one-dimensional foliation whose only sink is X and only source is X[1].","keywords":["cluster complex","2-Calabi-Yau category","X-evolution flow","X-foliation","green mutation","cluster exchange graph","homotopy type","quiver representations"],"falsifier":"Trace an $X$-leaf through a cell-crossing in the cluster category of a Euclidean quiver with $X$ in the vector-bundle component: if the leaf is a closed circle, Theorem 6.9's semi-compactness claim is false. More generally, if any nontrivial point other than $X$ or $X[1]$ has zero downward flow in a category satisfying Condition (*), the singularity analysis of Lemma 3.13 fails.","tokens_in":40202,"feed_emoji":"🌀","tokens_out":10823,"duration_ms":87927,"temperature":0.7,"pith_summary":"The paper introduces, for every point $X$ in the cluster complex of a 2-Calabi-Yau category, a global flow called the $X$-evolution flow. Its central claim is that this flow decomposes the whole cluster complex into one-dimensional pieces forming a piecewise linear foliation with exactly two singularities: the sink $X$ and the source $X[1]$. The construction is a continuous refinement of the discrete green mutation rule on the cluster exchange graph: for a generic $X$, the direction in which flow lines cross codimension-one faces is exactly the green direction. In the cluster categories of Dynkin and Euclidean quivers the foliation is compact or semi-compact for suitable $X$, and this determines the homotopy type of the cluster complex — a sphere in the Dynkin case and contractible in the Euclidean case — and implies that the fundamental group of the cluster exchange graph is generated by squares and pentagons.","feed_headline":"Every cluster point defines a foliation with exactly two singularities","feed_subtitle":"Continuous refinement of green mutation; yields homotopy type of cluster complexes for Dynkin and Euclidean quivers.","key_machinery":"The load-bearing object is the $X$-evolving triangle, a triangle $X \\to W \\to U \\to X[1]$ with $\\mathrm{Ext}^1(U,W)=0$ and no indecomposable summand shared between $W$ and $U$. For a cell $V$, the downward version takes $U$ to be the right minimal $V$-approximation of $X[1]$; the flow then sends a point in the cell along the vector $W-U$, or along the trivial rays to $X$ or $X[1]$. The key mechanism is that these local vectors are constant on parallel classes inside each cell and agree at codimension-one walls up to the required gluing condition, so local leaves assemble into global leaves. The irreducible-extension functors $\\Psi^{\\mathrm{down}}_X$ and $\\Psi^{\\mathrm{up}}_X$ give a trace that is constant on leaves and connects the foliation to Calabi-Yau reduction.","core_discovery":"Working in a 2-Calabi-Yau triangulated category $\\mathcal{C}$ that admits cluster-tilting sets and satisfies the condition that every rigid object is part of a partial cluster tilting set, the paper fixes a point $X$ in the cluster complex $\\mathrm{Cpx}(\\mathcal{C})$. Around each cell it forms the downward $X$-evolving triangle $X \\to W \\to U \\to X[1]$, where $U$ is the right minimal approximation of $X[1]$ by the cell, and defines the local flow direction $W-U$, with the two trivial branches $X-V$ and $V-X[1]$. These local directions are compatible across shared faces, so the leaves glue into a piecewise linear one-dimensional foliation with $X$ as its unique sink and $X[1]$ as its unique source. The paper proves this $X$-foliation exists for every point $X$, shows that the orientation of the exchange graph it induces is exactly green mutation when $X$ is generic, and proves compact or semi-compact behaviour in the Dynkin and Euclidean cases, leading to the homotopy-type and fundamental-group conclusions.","pith_inferences":["Editorial inference: since the green-mutation theorem only uses sign-coherence and tropical duality, the same flow construction should refine green mutations in skew-symmetrizable cluster algebras, not only in 2-Calabi-Yau categorifications.","Editorial inference: compactness of an $X$-foliation amounts to a deformation retraction of the cluster complex minus the source onto the sink, so the flow gives a visual certificate for the homotopy type; tracing representative leaves in finite examples is a direct way to test the conjectured compactness beyond Dynkin type.","Editorial inference: the leaf-wise invariant trace maps the cluster complex minus its two singularities to the Calabi-Yau reduction; in semi-compact cases this should yield a filtration of the complex by trace value, offering an inductive tool for homotopy computations in other tame or wild settings."],"forward_implications":["Every cluster complex carries, for each point $X$, a piecewise linear one-dimensional foliation with only two singularities, so the entire complex is a union of flow lines from $X[1]$ to $X$.","For a generic point $X$ in a top cell, the cell-crossings of the downward flow orient the unoriented cluster exchange graph exactly as green mutation with respect to $X[1]$, so green mutation is a discrete shadow of a continuous flow.","For a Dynkin quiver $Q$, the $X$-foliation is compact and induces $\\mathrm{Cpx}(\\mathcal{C}(Q)) \\simeq \\Sigma\\,\\mathrm{Cpx}(\\mathcal{C}(Q\\setminus 0))$, giving inductively that the cluster complex is homotopy equivalent to an $(n-1)$-sphere.","For a Euclidean quiver, the $X$-foliation is compact for rigid regular simple $X$ and semi-compact for the affine type A cases outside tubes, and the cluster complex is contractible.","For Dynkin and Euclidean quivers, the fundamental group of the cluster exchange graph is generated by squares and pentagons."],"supporting_citations":[{"why":"supplies the cluster-category setting and cluster-tilting sets in which Cpx(C) is defined","marker":"[BMRRT]"},{"why":"gives the index triangle for a top cell, which coincides with the X-evolving triangle for a cluster","marker":"[DK]"},{"why":"defines partial cluster tilting subcategories and mutation, the basic cells of Cpx(C)","marker":"[KR]"},{"why":"defines cluster complexes for marked surfaces, the homotopy-type problem the paper categorifies","marker":"[FST]"},{"why":"introduces green mutation, the discrete orientation rule that the X-flow refines","marker":"[Ke2]"},{"why":"provides Calabi-Yau reduction, used to identify Cpx(CzX) with Link(X)","marker":"[IY]"},{"why":"gives the fundamental-domain description of exchange graphs and the square/pentagon generating framework","marker":"[KQ2]"},{"why":"interprets green mutation as backward simple tilting, the categorical orientation recovered by the flow","marker":"[Qy2]"}],"fun_headline_variants":["Every cluster point sprouts an X-foliation with two singularities","X-evolution flow: a continuous generalization of green mutation","Cluster complexes foliate from any point, with one sink and one source","X-foliations on cluster complexes: compact for Dynkin, semi-compact for Euclidean","From green mutation to X-evolution: foliations with unique sink and source"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"Condition (*) — that every rigid object in $\\mathcal{C}$ is a partial cluster tilting set — is the load-bearing premise: it guarantees that $X$-evolving triangles exist and are unique for every cell, and the main theorems are only proved for categories with this finiteness property.","fun_headline_variants_meta":{"raw":{"variants":["Every cluster point sprouts an X-foliation with two singularities","X-evolution flow: a continuous generalization of green mutation","Cluster complexes foliate from any point, with one sink and one source","X-foliations on cluster complexes: compact for Dynkin, semi-compact for Euclidean","From green mutation to X-evolution: foliations with unique sink and source"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000937,"raw_usage":{"total_tokens":4025,"prompt_tokens":979,"completion_tokens":3046,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":595,"completion_tokens_details":{"reasoning_tokens":2949}},"tokens_in":595,"tokens_out":3046,"duration_ms":18832,"temperature":1.0,"reasoning_tokens":2949,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T13:58:01.861365+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Trace an $X$-leaf through a cell-crossing in the cluster category of a Euclidean quiver with $X$ in the vector-bundle component: if the leaf is a closed circle, Theorem 6.9's semi-compactness claim is false. More generally, if any nontrivial point other than $X$ or $X[1]$ has zero downward flow in a category satisfying Condition (*), the singularity analysis of Lemma 3.13 fails.","supporting_citations":[],"review_version":1}